The Digital Transformation Journey in BP Upstream...(BHGE and Palantir deploying solutions in a...

#PIWorld ©2019 OSIsoft, LLC

The Digital Transformation Journey in BP Upstream

Steve Beamer

1


Global Operations in BP

• Responsible for safe, compliant, reliable, and efficient operations

• We operate upstream production assets (onshore and offshore) and

midstream transportation and processing activities around the globe

Alask

a Gulf of Mexico Azerbaijan

2


3

Large and Diverse


4

The Challenges That Started Us on This Journey

• Need for a step-change improvement in Process Safety

• Need to improve competitiveness

• Recognition by Executive Team that it was hard for people to access information and this was driving suboptimal decision making

Engineers spending 80% of their time finding information, 20% of time troubleshooting

No central repositories, single owners or commonality of tools

• Massive amounts of data, disconnected from other related data and disconnected from end users/consumers

• Compartmentalized organization resulting in compartmentalized information


5

A “Simplified View” Everything is Related, Nothing is Connected

Many SME’s

work in a

Single

Database

focused on

their

expertise

(Inspection,

Maintenance

etc.)

Much of the

business value

comes from

relating the

data to other

data in

different

databases

(equipment

work orders to

plant

conditions)


6

Our Journey

• 2014 – Began migration of data to data lake to facilitate “Big Data” projects

Worked with Palantir, a tech startup company across several parts of the business • Based on successes in other parts of business, CEO sponsored global adoption

• Dedicated team with executive leadership eliminated barriers to speed deployment

This was seen as a step change in enabling multiple technology projects to succeed – Data access via historians is now considered to be “business critical” – no longer a nice to have.

• 2016 BHGE collaboration to build a “Plant Operations Advisor”. Problem – align Real Time (PI) tags to common hierarchy to feed to BP Data Lake to then feed to BHGE Data Lake.

Completed first asset ~17,000 tags, fed from our PI Historians took 6+ months to map and align with 5 FTE resources including documenting the process.

Realized their must be a better way to do this and asked for help from OSIsoft


7

The Accelerator Element AssetHub: Connect, Manage, and Share Asset Data


8

Where We Are Today – 18 months later

• Total Assets: 33 ~ 70% complete – taking 4 – 6 weeks per asset with 1 FTE

• Time series tags: 1m+. Data connections built 30m+.

• Other data sources: 6000+ (i.e. SAP, Hazops, Spreadsheets, limits etc.). For example, we are now automatically storing a history of HAZOP demands which will support better analysis in the future

• Annual data points made accessible via “digital twin” models: 100B+ (mostly time series data)

• Alignment of internal data sources across multiple systems in the Data Lake vs attempting to do so in the Core Systems. Availability of data, combined with live analysis to assist decision making is now tangible.


9

Setting the Foundations for Scaling


10

What Enterprise Framework Enabled (PI-Vision on AF)

• Common data view (digital twin) allowed us quickly build and deploy AF’s

• Standardized process analysis template for global adoption

• One global workstream instead of multiple assets/regions “reinventing the wheel”


11

Global Templates for Analytics


12

PI-AF Vision – One Team delivers Analysis in 3 days!

• During the PI-AF roll out workshop in the North Sea, the Glen Lyon(GL)

Team presented a business problem with produced water Progressive

Cavity Pumps which were significantly impacting production. The pumps

supported a 20 mboed production improvement opportunity. (Approx

$400m/yr)

• The Analysis provides absolute and theoretical values of motor and

hydraulic power and efficiency and is in the process of being extended to

show leakage flows (also known as slip) and power offset relative to the

Manufacturer’s curves for the pumps. The analytic provides Absolute &

Theoretical values which are visualised to create a clear insights into any

potential deteriorating performance. Work is already underway to develop

the Analysis even further.

• The PI-AF monitoring capability is part of a suite of Analytics and

Dashboards to be deployed to additional regions over the coming months.

Feb 2018

Remote Operations Pump Analytic delivered to Glen Lyon ACE for Critical Pump Start-Up.


13

How Are We Measuring Success?

• Business Improvement – Availability/Reliability of Plant, Production Improvement. Early guesstimates used to justify the project. Actuals against baseline is measured.

• Speed of Development and Deployment – Metrics developed in Agile process in tranches of 6 and rapidly deployed globally

• End user feedback – both quantitative (usage of the tools and analytics) and qualitative (regular engagement)


14

Learnings • No need to change core systems – maintain your investment AND when the time

comes to change them – modern architecture and connectivity are extremely important for the future. VALUE CONNECTIVITY OVER FEATURES in selecting tools in the future.

• Data Lake is the beginning not the end. By itself, it does not solve any problems.

• Build useful data sets (digital twins) by equipment class for multiple use cases vs a use case per application.

• AF templates help to deploy at pace and enabled other initiatives as they materialized. Hard to scale AF without a “Federal Structure”.

• Enable the data owners to curate and maintain the models in order to democratize the data.

• We had to work with our key suppliers as partners vs a traditional customer/supplier relationship. MVP (Minimum Viable Product) vs POC (Proof of Concept). MVP Scale.


15

Summary

• Early collaboration with Strategic Partners (Palantir, BHGE, Microsoft) put is in

the cloud/big data world early. Adoption of OSIsoft Enterprise agreement

enabled the business to support many of the big data projects.

• Collaborating with key suppliers helped us to scale (MVP Scale)

(BHGE and Palantir deploying solutions in a cloud/big data environment)

(OSIsoft and Element Analytics on organizing our data and deploying AF quickly)

• Alignment of data to a common Federal Structure in the Data Lake allowed us

to develop and deploy fast for PI-Vision, POA, and several recent initiatives

through the use of AF. Data Lake is one enabler not the complete answer.

• Use cases we did not anticipate at the time are leveraging the same data

models to move more quickly.


Questions?

Please wait for

the microphone

State your

name & company

Please remember

DOWNLOAD THE MOBILE APP

16


17


BACKUP


Element AssetHub

Publish Models

Data Tables

Data Export Portal Enterprise

Calculations

Asset

Attributes

Asset

Templates

Hierarchical

Views

Build Operate

Infrastructure

Foundation

Architectural Design

Connect Manage Share Design Maintain

Map Assets

Asset

Pipelines

Data

Integrity

Checks

Model

Integrity

Perpetual

Data

Master Data

Data Agents

Connectivity APIs

Data Import Portal

Agent Ingress

Element Graph

Big Compute Container-

based

Managed SaaS

Self-healing

Elasticity

Transform

Sources

Enterprise Security

Authentication

Integration

Encryption

Data

Integrity

Report

The Digital Transformation Journey in BP Upstream...(BHGE and Palantir deploying solutions in a...

Documents

Transcript of The Digital Transformation Journey in BP Upstream...(BHGE and Palantir deploying solutions in a...